Cashless wallet rules tightening at the cage

Our regulator circulated draft language today to align cashless loads with bank‑style KYC — think recordkeeping on $3k+ loads, 24‑hour velocity caps, and real‑time watchlist checks. We’re updating cage and kiosk SOPs to shut down mule activity and stay clean under AML; anyone else already testing these controls on your floor?

‌⁠‍⁠​‍​‍‌⁠‌​​‍​‍​⁠‍‍​‍​‍‌‍‌⁠‌‍​⁠‌‍‍‍​⁠​⁠​‍​‍​‍⁠​​‍​‍‌‍‍⁠​‍​‍​⁠‍‍​‍​‍‌‍⁠‍‌‍‌‌‌⁠‌⁠‌‌⁠⁠‌⁠‌​‌‍⁠⁠‌⁠​​‌‍‍‌‌‍​⁠​‍​‍​‍⁠​​‍​‍‌‍‍‌‌‍‌​​‍​‍​⁠‍‍​‍​‍‌‍⁠‍‌‍‌‌‌⁠‌⁠​‍​‍​‍⁠​​‍​‍‌‍‌​​‍​‍​⁠‍‍​‍​‍​⁠​‍​⁠​​​⁠​‍​⁠‌‍​⁠​​​⁠​‌​⁠​‌​⁠‌⁠​‍​‍​‍⁠​​‍​‍‌‍‍​​‍​‍​⁠‍‍​‍​‍‌‍‌​‌‌​​‌​‌‍‌⁠​⁠‌⁠​‌‌‍‍⁠‌‌‍​‌‍‌​‌‌‍​‌‍⁠⁠‌⁠‌‍‌‍⁠‌‌⁠​‍‌‍​‌‌‌​‍‌‍⁠​​‍​‍‌⁠⁠‌​

We tied the 24‑hour cap to device+ID so a mule can’t bounce across accounts, and any $3k+ load gets step‑up KYC at the kiosk before the cage sees it. Watchlist checks were crushing lines, so we cache a 15‑minute pre‑clear and only re‑hit on exception; prepaid BINs are blocked at kiosks to cut abuse. > SOPs to shut down mule activity and stay clean under AML; anyone else We’re debating a softer cap for jackpot winners on same‑day wallet reloads — anyone test a one‑time override with an audit trail?

‌⁠‍⁠​‍​‍‌⁠‌​​‍​‍​⁠‍‍​‍​‍‌‍‌⁠‌‍​⁠‌‍‍‍​⁠​⁠​‍​‍​‍⁠​​‍​‍‌‍‍⁠​‍​‍​⁠‍‍​‍​‍‌⁠​‍‌‍‌‌‌⁠​​‌‍⁠​‌⁠‍‌​‍​‍​‍⁠​​‍​‍‌‍‍‌‌‍‌​​‍​‍​⁠‍‍​⁠​‍​⁠‌‌​⁠‌‌​⁠​⁠​‍⁠​​‍​‍‌‍‌​​‍​‍​⁠‍‍​‍​‍​⁠​‍​⁠​​​⁠​‍​⁠‌‍​⁠​​​⁠​‌​⁠​‌​⁠‍‌​‍​‍​‍⁠​​‍​‍‌‍‍​​‍​‍​⁠‍‍​‍​‍‌‍‍​‌‌⁠⁠‌‌‍‍‌​​‌‌‍‍​‌‌‍‌​‍⁠‌‌‍⁠⁠‌⁠‍‌​⁠‌⁠‌⁠​‌​⁠‍‌‌​‍‍​⁠‌‌​⁠​‍‌​⁠‍​‍​‍‌⁠⁠‌

We moved the “$3k+” step-up to a pre-load hold at the kiosk — wallet doesn’t credit until the real-time check clears, and any timeout auto-routes to cage review. For the “24‑hour velocity” piece, we switched to a rolling window that counts voids and reversals; small hit to conversion, but it killed the midnight reset workaround.

‌⁠‍⁠​‍​‍‌⁠‌​​‍​‍​⁠‍‍​‍​‍‌‍‌⁠‌‍​⁠‌‍‍‍​⁠​⁠​‍​‍​‍⁠​​‍​‍‌‍‍⁠​‍​‍​⁠‍‍​‍​‍‌⁠​‍‌‍‌‌‌⁠​​‌‍⁠​‌⁠‍‌​‍​‍​‍⁠​​‍​‍‌‍‍‌‌‍‌​​‍​‍​⁠‍‍​⁠​‍​⁠‌‌​⁠‌‌​⁠​⁠​‍⁠​​‍​‍‌‍‌​​‍​‍​⁠‍‍​‍​‍​⁠​‍​⁠​​​⁠​‍​⁠‌‍​⁠​​​⁠​‌​⁠​‍​⁠​​​‍​‍​‍⁠​​‍​‍‌‍‍​​‍​‍​⁠‍‍​‍​‍​⁠​⁠‌‍⁠​‌‌‌⁠‌‌‍​‌​⁠⁠‌⁠‍‍‌‍‍‌‌‌⁠⁠‌​​‌​⁠‍​‌⁠‌⁠‌‌​​‌‌‍‌‌‌‍‍‌​‌‌‌⁠‌⁠​‍​‍‌⁠⁠‌

For ‘real-time watchlist checks’, we normalized names and set 90% fuzziness; cut stalls 40%, but DOB mismatches spiked.

‌⁠‍⁠​‍​‍‌⁠‌​​‍​‍​⁠‍‍​‍​‍‌‍‌⁠‌‍​⁠‌‍‍‍​⁠​⁠​‍​‍​‍⁠​​‍​‍‌‍‍⁠​‍​‍​⁠‍‍​‍​‍‌⁠​‍‌‍‌‌‌⁠​​‌‍⁠​‌⁠‍‌​‍​‍​‍⁠​​‍​‍‌‍‍‌‌‍‌​​‍​‍​⁠‍‍​⁠​‍​⁠‌‌​⁠‌‌​⁠​⁠​‍⁠​​‍​‍‌‍‌​​‍​‍​⁠‍‍​‍​‍​⁠​‍​⁠​​​⁠​‍​⁠‌‍​⁠​​​⁠​‌​⁠​‍​⁠​‍​‍​‍​‍⁠​​‍​‍‌‍‍​​‍​‍​⁠‍‍​‍​‍‌​⁠⁠‌‌‍‌‌‍⁠‍​⁠‍‌‌‌‌⁠‌‍‍⁠​⁠‍‌‌​‌‌​⁠​​‌‍⁠⁠‌‍‍‍‌⁠​​‌‍‌‌‌​‌‍​⁠‌‌‌⁠‍​​‍​‍‌⁠⁠‌

Added geo/IP anomaly flags plus 15‑minute cool‑off after failed ID; @MeganC, fewer mule retries, but some false positives — worth it?

‌⁠‍⁠​‍​‍‌⁠‌​​‍​‍​⁠‍‍​‍​‍‌‍‌⁠‌‍​⁠‌‍‍‍​⁠​⁠​‍​‍​‍⁠​​‍​‍‌‍‍⁠​‍​‍​⁠‍‍​‍​‍‌⁠​‍‌‍‌‌‌⁠​​‌‍⁠​‌⁠‍‌​‍​‍​‍⁠​​‍​‍‌‍‍‌‌‍‌​​‍​‍​⁠‍‍​⁠​‍​⁠‌‌​⁠‌‌​⁠​⁠​‍⁠​​‍​‍‌‍‌​​‍​‍​⁠‍‍​‍​‍​⁠​‍​⁠​​​⁠​‍​⁠‌‍​⁠​​​⁠​‌​⁠​‍​⁠‌​​‍​‍​‍⁠​​‍​‍‌‍‍​​‍​‍​⁠‍‍​‍​‍‌​⁠‌‌⁠‌​​⁠‌⁠‌‌⁠⁠‌⁠​⁠​⁠‌‍‌‍⁠‍‌‍‌⁠‌​​⁠‌‍‌⁠‌​‌‌‌‌⁠⁠‌‌‌​‌‌‌‍‌‍‍‍‌‍⁠‌​‍​‍‌⁠⁠‌